IDE Controller

[LunarMist]

Does it really make any difference between the Masters, Slaves, Primary and Secondary of the main IDE controllers when only one device is used at one time and none are ootable? I think it's all meaningless and connect the cables whichever way. If not so what's the explanation?

Thank you.

 

[P5-133XL]

Yes it matters. The last plug should always be populated otherwise the cable is unterminated causing signal corruption.

Some controllers/motherboards are picky about the Master/Slave/Single (WDC) jumpers causing either slow detection or failure in detection of the drives.

 

[zx]

The last plug on the cble should be used for the master drive and the other for the slave.

However, it makes very little difference it most cables.

 

[LunarMist]

Of course I knew there was some reason but forgot everything about IDE since the days before the SR crash. One of the two channels has a master only, so it is OK. The problem is that that the other channel has one or two drives. The slave drive is permanently connected, but the master is connected to the removable drive tray which is always powered up. The removable drive could be Maxtor, Seagate or Western Digitalis. He tried the WD drive as the internal slave or master with no slave and had non-start or crashing probelms as you surmised.

I cannot configure the removable drive as Slave because it must also be used on other bays in other computers, configured to Master, including the ADS Pyro. There is no data loss -- I must be lucky so far, but this is not good. I still have six unused ports on the motherboad (two IDE and two SATA), but they always ask for the driver every boot. Even with newest drivers it still happens every time, so is too anoying to tolerate. Some say I should upgrade to the XP, but then some of my old software and hardware will not work, so there are many hundreds more dollars flushed.

(Sorry, it was not intended to be a rant, but the computers and OS and hardware and applications are just hopless to continue upgrading.)

 

[LunarMist]

Er, "Digitalis." :?: Maybe I need some digitalis for my heart after this, but I mean Western Digital.

 

[jtr1962]

The master device must be on the end of the cable if the controller is ATA66 or better(which means it uses the special 80-pin cable). If it is an older machine(either ATA-33 or DMA-2) then it doesn't matter, although I still follow that practice out of habit. In fact, I even use 80-pin cables on such machines as it represents a more reliable data path(and I have a bunch of them anway from hard drive purchases and eBay purchases of Promise controllers).

 

[LunarMist]

Of course the Master is always at the end of the cable if it is attached. Now I use the add-in ATA/133 controller because of the issues with the ATA/100 controllers of the ICH5R (i875P) and the Promise SATA-ATA/133.

 

[Fushigi]

Suggestions, some of which will cost:

- Put the slave drive in a tray and dismount it when no master is present. If your access needs support that. Or flip it between master & slave and insert it into whichever slot is appropriate.

- Put the slave in a USB2/Firewire enclosure and remove it from the chassis altogether. So the internal cable is either unoccupied or occupied by a master only.

- Add another IDE cable/chain to the machine and make the current slave the master of the new chain. Keep the tray as master & have it be the only device on the chain.

- Convert the trays to external enclosures and ditch the trays.

- If you don't really need ATA66+ performance/capabilities, drop in an ATA33 controller and run in that mode, which, from what jtr said, would enable you to work as-is.

 

[LunarMist]

I may buy Windows XP as it appears to be a reasonably priced option to solve dilemmas. (Sorry if this one is turning into a Tech Support thread, I did not intend it.)